---
pretty_name: TTM4HVAC – Training dataset (source-all) 
tags:
- ttm4hvac
- hvac
- time-series
- energy
task_categories:
- time-series-forecasting
papers:
- title: "Transfer learning of building dynamics digital twin for HVAC control with Time-series Foundation Model"
  url: https://arxiv.org/abs/XXXX.XXXXX
  authors: "Ferran Aran Domingo"
license: mit
---

# TTM4HVAC – Training dataset (source-all) 

This dataset contains HVAC and weather time-series data used to train the **source-all** TinyTimeMixer model (`gft/ttm4hvac`), the main model of the TTM4HVAC project.

It aggregates all available source-building data under *default* and *non-default* conditions.

Check out the paper [arXiv:XXXX.XXXXX]() (to be released) and visit the main repository [ttm4hvac](https://huggingface.co/gft/ttm4hvac) for further details.

## Columns

- `time`
- `Outdoor Air Temperature (C)`
- `Heating Setpoint (C)`
- `Cooling Setpoint (C)`
- `Room Air Temperature (C)`
- `Outdoor Humidity (%)`
- `Wind Speed (m/s)`
- `Direct Solar Radiation (W/m^2)`
- `HVAC Power Consumption (W)`
- `series_id`
- `is_default`

## Usage

```python
from datasets import load_dataset

ds = load_dataset("gft/ttm4hvac-source-all-train")
df = ds["train"].to_pandas()
```

# ✒️ Citation

If you use this model or datasets, please cite:

```
**F. Aran**,  
*Transfer learning of building dynamics digital twin for HVAC control with Time-series Foundation Model*,  
arXiv:XXXX.XXXXX, 2025.  
https://arxiv.org/abs/XXXX.XXXXX
```